921B triangle again. Success!

[jorgen.bergfors at idg.se]

Yesterday evening I managed to get an almost perfect triangle wave out of (my 
copy of) the Moog 921B sawtooth to triangle waveshaper!
To get the wave symmtrical, I had to change the DC offset of the input signal. 
Strange I didn't think of this at first. I feed it a signal from the Doepfer 
PMS VCO and that is hardly identical to te Moog one (my Doepfer saw goes from 
5,3 to 10 volts peak to peak). With two extra resistors I can get the symmetry 
just the way I want it.
 To reduce the glitch at the peak of the wave, I added a 1n capacitor in series 
with a 8k02 resistor between the base and emitter of the first tranny. Don't 
ask me why this work. It creates an upward spike, that cancels out the downward 
spike in the waveform. The canellation isn't perfect at all frequencies, but it 
is a clear improvement. Over 4 kHz the waveform gets sligthly distorted but 
nothing serious. It is still much better than without the addition. The 
capacitors that are present in the Moog design (at the output) can be reduced. 
I removed the 1n one and reduced the 5n one to 1n. This way the amplitude isn't 
reduced at high frequencies.
I had to change some of the other resistors, to adapt the circuit to +-15 volts 
supply.
To be able to fine tune the continuity of the waveform at the positive peak, I 
added a 500 ohm trimmer in seres with the input resistor (reduced to 1k). 
There you have it -- a perfectly good saw to triangle converter using only two 
transistors, two capacitors and a bunch of resistors. Now I'm going to tackle 
the triangle to sine converter from the 921B.

/Jorgen

P.S. Contact me if you are interested in building this circuit. Schematics are 
available.